Research on personality disorders is plagued by methodological difficulties. The development of structured interviews for personality disorders has addressed concerns regarding interrater reliability, but even with structured interviews, there is relatively little evidence of the clinical and predictive validity of the diagnoses made. The most pressing scientific issue at present is documenting the validity of diagnoses derived from such measures, and the primary goal of the present proposal is to collect data necessary to address the validity question. The first specific aim is to conduct a study of the clinical and predictive validity of the Personality Disorder Examination (PDE) and the Structured Interview for the DSM-III Personality Disorders (SIDP) in a sample of 200 patients with nonpsychotic diagnoses on Axis I of DSM-III-R. The method for establishing validity has been pretested in a pilot study and relies heavily on Spitzer's recommendations for the use of the "LEAD standard." The acronym LEAD refers to three key concepts: the use of Longitudinal information, ratings of Expert clinicians, and the collection of All Data available. The second aim is to compare the diagnoses derived from clinical consensus and the structured interviews (PDE and SIDP) with those generated by the Wisconsin Personality Inventory (WISPI), the most promising self-report instrument currently available for Axis II disorders. The third aim is to investigate the influence of symptomatic state on what are assumed to be trait assessments done by these different methods (clinical consensus, structured interview, and self-report). The "state vs. trait" issue (especially as it influences the concordance between informants and among evaluation methods) is a thorny but fundamental problem in the attempt to establish valid assessment procedures. A longitudinal design will allow us to measure the stability of trait assessments (done using different methods and relying on different informants) as the patient's symptomatology varies over a 12-month period.